Classifications

• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
  • G01B11/00—Measuring arrangements characterised by the use of optical techniques
  • G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
  • G01B11/2441—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures using interferometry
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
  • G01B11/00—Measuring arrangements characterised by the use of optical techniques
  • G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
  • G01B11/25—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures by projecting a pattern, e.g. one or more lines, moiré fringes on the object
  • G01B11/2518—Projection by scanning of the object
  • G01B11/2527—Projection by scanning of the object with phase change by in-plane movement of the patern

Definitions

• the invention relates generally to the measurement of surface contours and more particularly to a non-contact interferometric system and method for the measurement of surface contours.
• Interferometric surface measurement systems have been developed which permit measurements of the surface of an object without physical contact.
• Coherent optical sources are used to generate a fringe pattern on the surface of the object and a camera acquires images of the fringes on the surface for analysis
• a diffraction grating is positioned in the path of a laser beam to generate multiple coherent laser beams at various angles to the original beam path.
• a focusing objective and spatial filter are used to isolate the desired diffracted beam pair.
• One or more additional diffraction gratings are utilized to project at least one additional set of fringes onto the object surface. This multiplexing of different gratings into the beam path poses many challenges.
• Moving different gratings into the beam path and shifting each grating to implement phase shifts generally requires multiple mechanical components that add weight, size, complexity and cost to the system.
• the frequent movement of components affects the stability and therefore the accuracy of the measurement data.
• measuring the displacement of a diffraction grating during the phase shift process with sufficient precision and accuracy can require expensive measurement components such as capacitance gauges.
• Noise sources also typically limit the measurement data.
• a high-contrast, granular speckle pattern is typically observed. Speckle results in part from the roughness of the object surface.
• the microscopic roughness of the surface contributes randomly phased contributions of the scattered laser light. These contributions interfere with one another to produce complex intensity variations across the surface of the object as viewed from a distance. Speckle introduces fine-scale intensity fluctuations (i.e., intensity noise) in the observed fringe pattern on the object surface. Shot noise contributions from the individual detectors in the camera can further limit the accuracy of measurement data.
• the invention features a method for determining a fringe number for a pixel for an offset channel in a multiple channel interferometric surface contour measurement system.
• a measured fringe number is determined for the offset channel and adjusted by a collinear adjustment value to generate a collinear fringe number.
• the collinear fringe number is unwrapped to determine an integer portion and adjusted by the collinear adjustment value to obtain the fringe number for the offset channel.
• the invention features a method for determining a fringe number for a plurality of pixels for an offset channel in a multiple channel interferometric surface contour measurement system.
• a measured fringe number is determined for the offset channel for each pixel.
• Each measured fringe number is adjusted by a collinear adjustment value for the respective pixel to obtain a collinear fringe number for the respective pixel.
• the collinear adjustment value is determined in part from a plurality of predetermined linear coefficients for the respective pixel.
• the collinear fringe numbers are unwrapped to determine an integer portion of the fringe number for each pixel.
• the unwrapped collinear fringe numbers are then adjusted by the collinear adjustment value of the respective pixels to obtain the fringe numbers.
• the invention features a multiple channel interferometric surface contour measurement system.
• the measurement system includes a multiple channel interferometer projector, a digital camera and a processor.
• the processor communicates with the multiple channel interferometer projector and the digital camera.
• the multiple channel interferometer projector projects fringe patterns onto a surface of an object and includes an offset channel. Images of the projected fringe patterns on the surface of the object are acquired by the digital camera.
• the processor determines, for each of a plurality of pixels, a measured fringe number for the offset channel in response to images received from the digital camera.
• the processor adjusts the measured fringe numbers to generate a collinear fringe number for each pixel, unwraps the collinear fringe numbers and adjusts the unwrapped fringe numbers to obtain the fringe number for each pixel.
• the invention features a method for determining a fringe number for a point on an object surface.
• a first fringe pattern having a spatial frequency associated with a first source separation distance is projected onto the object surface and a second fringe pattern having a spatial frequency associated with a second source separation distance is projected onto the object surface.
• the first and second source separation distances are integer multiples mi and mi, respectively, of a common base distance.
• the integer multiples mj and mj are relatively prime and greater than an integer fringe number for the point for the respective fringe pattern.
• Integer portions N(mi)j and N (JM) 2 of the fringe numbers for the point for the first and second fringe patterns, respectively, are dete ⁇ nined in response to the integer multiples mi and mj.
• FIG. 1 is a diagram of a surface contour measurement system in which the surface of an object is irradiated with structured light patterns generated according to interferometric principles.
• FIG. 2 is another view of the system of FIG. 1 and includes a magnified view of a surface contour of the object.
• FIG. 3 depicts a portion of an image of the surface of the object of FIG. 2 as irradiated by a first fringe pattern.
• FIG. 4 shows a view of a portion of an image of the surface of the object of FIG. 2 as irradiated by a second fringe pattern.
• FIG. 5 A and FIG. 5B illustrate front and rear perspective views, respectively, of a multiple channel interferometer projector according to an embodiment of the invention.
• FIG. 6 A and FIG. 6B illustrate exploded perspective views of the projector shown in FIG. 5A and FIG. 5B, respectively.
• FIG. 7 shows a perspective view of a diffraction grating module of the projector of FIG. 6 A and FIG. 6B according to an embodiment of the invention.
• FIG. 8A and FIG. 8B illustrate rear and front perspective views, respectively, of the diffraction grating module of FIG. 7 with the grating substrate removed.
• FIG. 9 illustrates an exploded perspective view of a multiple channel interferometer projector having an intensity shaping module according to another embodiment of the invention.
• FIG. 10 is a cross-sectional diagram of a crossbar for rigidly coupling a camera to a projector according to an embodiment of the invention.
• FIG. 11 is a perspective view of a crossbar for rigidly coupling a camera to a projector according to an embodiment of the invention.
• FIG. 12 depicts the image of FIG. 3 where phase is expressed in terms of the fringe number.
• FIG. 13 illustrates an end view of a multiple channel interferometer projector with three active interferometer channels.
• FIG. 14 is a flowchart representation of a method to calculate the fringe numbers for a pixel for fringe patterns generated by a multiple channel interferometer projector according to an embodiment of the invention.
• the present invention relates to a multiple channel interferometric surface contour measurement system.
• the measurement system includes a multiple channel interferometer projector, a digital camera and a processor. Fringe patterns generated by spatially separate channels in the projector are projected onto an object surface to be measured.
• the digital camera acquires images of the fringe patterns and the processor determines surface contour data from the fringe patterns.
• fringe numbers are determined for points on the object surface based on image data.
• the fringe numbers are modified according to collinear adjustment values so that the modified fringe numbers correspond to a common, collinear axis for the interferometer projector. After unwrapping the modified fringe numbers, the unwrapped values are modified by the collinear adjustment values to obtain accurate fringe numbers for the pixels in each interferometer channel.
• FIG. 1 is an illustration of a surface measurement system 10 in which a projector 14 irradiates a surface of an object of interest 18 with structured light patterns that are generated based on interferometric principles. Images of the irradiated object surface 34 are acquired by a camera 22 that is rigidly coupled to the projector 14 via a crossbar 26. The camera 22 is electronically coupled to a processor 30 which is configured to process the acquired images. The processor 30 can be coupled to the projector 14 to control various aspects of the structured light patterns.
• the surface measurement system 10 is configured to determine a three- dimensional profile of the surface contour 34.
• the processor 30 calculates relative coordinates in three dimensional space for a significant number of points on the object surface 34.
• a coordinate system 38 is shown in FIG. 1 to provide a reference for the surface contour measurements. They-axis of the coordinate system 38 is perpendicular to the plane of the figure.
• An image plane 42 e.g., a detector array
• the optical axis 44 of the camera 22 (which is perpendicular to the image plane 42) is parallel to the z-axis of the coordinate system 38.
• the x-y coordinates of a point P in the camera field of view generally are directly determinable from the known geometry and optical parameters of the camera 22; however, a z-coordinate Zp of the point P is not directly determinable.
• Structured light patterns are used to calculate the z-coordinates of points on the object surface 34.
• Images of structured light patterns irradiating the object surface 34 provide information from which the processor 30 can calculate distances Zp between points P on the object surface and the camera image plane 42.
• Such calculations are based on triangulatioii, using a known distance L between the camera 22 and the projector 14, and a known angle ⁇ between a central axis 46 of the projector 14 and an axis of the crossbar 26.
• a coherent light source e.g., a laser
• the two beams are directed toward the object surface 34 and caused to interfere with each other so as to generate a pattern of interference fringes comprising periodic bright lines separated by dark lines.
• the process of separating light from a coherent light source into two beams and causing the two beams to interfere to generate a fringe pattern is referred to as "single channel interferometry.”
• FIG. 2 is an illustration of the surface measurement system 10 of FIG. 1 and includes a magnified view of the surface contour 34 of the object of interest 18 around a point of interest P for which the distance Zp is to be determined.
• the projector 14 is configured to irradiate the object surface 34 with a fringe pattern 50 depicted as a periodic wave.
• the peaks of the periodic wave represent the highest intensities and the troughs of the periodic wave represent the lowest intensities.
• the areas around the peaks represent bright lines or "fringes" on the object surface 34 that can be observed by the camera 22 and the areas around the troughs represent dark lines between neighboring fringes.
• the spatial frequency (i.e., periodicity) of the fringe pattern 50 generally is a function of the configuration of the projector 14 and the arrangement of two coherent light sources 14A and 14B or, equivalently, the apparent sources of two spatially distinct beams from a single coherent light source, that interfere with each other.
• the sources 14A and 14B are separated by a distance a (hereafter the "source separation distance") along an axis 48 perpendicular to the central axis 46 of the projector 14.
• the spatial frequency of the fringe pattern 50 is determined in part by the source separation distance a.
• the spatial frequency decreases (i.e., the fringes become "coarser") as the source separation distance a decreases.
• a single channel interferometer projector transmits a laser beam through a diffraction grating having a pitch d (i.e., the spacing between adjacent light perturbing components of the diffraction grating).
• the laser beam is divided into pairs of coherent beams propagating at various deflection angles from the grating according to their order and the pitch d of the grating.
• a portion of the laser beam passes through the diffraction grating undeflected as a zero order beam.
• the pitch d increases (i.e., as the grating becomes coarser)
• the deflection angles of the non-zero order beams decrease.
• the single channel interferometer utilizes a focusing objective and a spatial filter to isolate the pair of first order beams propagating from the diffraction grating and thereby provide two coherent beams required for interference.
• the focusing objective directs the two first order beams towards two apertures spaced apart at the source separation distance a in the spatial filter.
• This source separation distance a is determined from the known angles at which the first order beams exit the diffraction grating and the optical properties of the focusing objective. Accordingly, the spatial frequency of the fringe pattern 50 generated by the single channel interferometer is based on the diffraction grating pitch d.
• the fringe pattern 50 includes a central fringe having a fringe number N 0 and an intensity peak 54, and is associated with a point P 0 on the object surface 34 which is intersected by the central axis 46 of the projector 14. As illustrated, the coherent light sources 14A and 14B are disposed symmetrically about the central axis 46.
• the peak 54 of the central fringe N 0 serves as a reference point for the fringe pattern 50, on either side of which are other fringes consecutively labeled N+j, N + 2, ... moving to the right in the figure and N.j, N-2, ... moving to the left in the figure. Accordingly, the peak 54 of the central fringe N 0 is associated with a reference phase of 0° for the fringe pattern 50 and the phase between adjacent peaks, or adjacent troughs, is 360° or 2 ⁇ radians.
• the determination of the distance z P between the camera 22 and a point of interest P can be determined from the phase ⁇ between point Po at the intensity peak 54 and point P at an intensity 58 on the fringe N.;.
• the quantity N hereafter referred to as the "fringe number,” is given by
• ⁇ is the source separation distance described above
• ⁇ is the optical wavelength of the sources 14A and 14B
• x is the physical distance along the x-axis between the points P and P 0
• R is the physical distance between the projector 14 and the point P 0 as measured along the central axis 46 of the projector 14 from the midpoint of the sources 14A and 14B and the point P 0 -
• the equation above is generally valid provided that the source separation distance a and the distance x are relatively small compared to the distance R.
• observing the phase ⁇ in terms of the fringe number N from the acquired images of the fringe pattern 50 provides information relating to the parameters x and R which relate to the angle ⁇ and allow a determination of the distance z P by triangulation.
• the accuracy with which the distance z P can be determined is related in part to the spatial frequency of the fringe pattern 50.
• Profile changes in the surface contour 34 along the z direction correspond to changes in observed intensity of the fringe pattern 50.
• the resolution of the surface contour measurement increases as the change in intensity for a given change in distance along the z direction increases. Stated differently, as the slope of the individual fringes of the fringe pattern 50 increases, more accurate measurements of the distance z P can be made. This situation corresponds to "finer" fringes or, equivalently, an increased spatial frequency of the fringe pattern 50 on the object surface 34.
• the interferometric fringe pattern 50 has an infinite field of focus relative to the projector 14, images of the projected fringe pattern 50 are limited by practical constraints. For example, imaging a fine fringe pattern over a large range in the z direction requires an imaging system having a large depth of focus and high spatial resolution. These requirements are generally difficult to achieve simultaneously and a particular application typically requires a tradeoff between the two parameters. Accordingly, in some projector implementations, the grating pitch d and the source separation distance a for a given single channel interferometer are selected based, at least in part, on a compromise between high measurement resolution and camera depth of field.
• FIG. 3 depicts an image 62 of the object surface acquired by the camera.
• Image data is obtained using a two-dimensional matrix of pixels in the camera image plane.
• Each pixel is associated with an intensity value /based on the light scattered from the object surface and incident on the pixel,
• One such pixel 66 is shown as containing the point of interest P.
• the image 62 includes the alternating bright and dark lines of the fringe pattern in which the fringes are indicated with shading and are marked for reference with the designations N + s to Ns moving from left to right in the figure. It should be appreciated that the intensity values of pixels in a row of the image 62 reveal the continuous wave nature of the fringe pattern 50 shown in FIG. 2.
• the distance Zp to the point P can be calculated. Additionally, the x andy coordinates of the point P can be determined from knowledge of the geometry and optical properties of the camera. Performing the analysis and three- dimensional coordinate calculation for each pixel in the image 62 allows a detailed profile, or map, of the object surface to be determined.
• the phase ⁇ is calculated by (i) determining the "fractional phase" r/y rac of the intensity detected for the point P relative to its nearest maximum intensity peak, that is, the center of the fringe Ni and (ii) identifying in the image 62 the location of the zero degree reference phase corresponding to the intensity peak 54 in FIG. 2 so as to determine the portion of the total phase ⁇ that corresponds to an integer number of fringe peaks.
• a known phase shift technique can be utilized to determine the fractional phase ⁇ f r ⁇ c .
• the projector 14 is configured so that the fringe pattern 50 can be phase-shifted by two or more known values relative to its nominal position.
• the projector 14 can be configured to shift the fringe pattern 50 by shifting the diffraction grating so that the maximum intensity peak 54 of the center fringe N 0 is shifted either to the right or left of the point P 0 by a known amount (e.g., ⁇ 120°, or ).
• the fractional phase ⁇ pf rac can be unambiguously calculated if an image of the fringe pattern 50 is acquired at each of three different known phase shifts (e.g., +120°, 0° and -120°).
• the intensity / detected by the pixel 66 is determined (e.g., i +120 , Io and I -120 ). While phase shifts of ⁇ 120° and the reference position of 0° are used in the present example, it should be appreciated that other numbers of phase shifts and other values of phase shift can be utilized in the technique.
• the location of the zero degree reference phase in the image 62 is determined so that the portion of the total phase ⁇ that includes the integer number of fringes can be determined.
• the zero degree reference phase is indicated in the image 62, the corresponding center fringe N 0 is otherwise not readily apparent as the fringes N +3 to N. 3 appear identical.
• This fringe similarity results in a "2 ⁇ ambiguity because there is no analysis of a single image that yields the integer number of fringe peaks from the 0° phase fringe and the point P. Resolving the 2 ⁇ ambiguity to determine the integer number of peaks contributing to the phase ⁇ is referred to in the art as "fringe unwrapping.”
• Fringe unwrapping can be described with reference to FIG. 4 in which an image 62' of a different fringe pattern projected onto the object surface is acquired.
• the projector irradiates the object surface with a coarse (i.e., lower spatial frequency) fringe pattern.
• the coarse fringe pattern includes only one fringe within the camera field of view, namely the central fringe N 0 , wherein the fringe pattern is calibrated to have the same zero degree reference phase as the fringe pattern 50 shown in FIG. 2 and FIG. 3.
• the phase ⁇ corresponding to the point of interest P can be determined directly using the phase shift technique discussed above. Ambiguity is avoided because the phase ⁇ variation across the image 62' is less than 2 ⁇ .
• an iterative technique can be employed in which the projector is configured to alternately project coarse and fine fringe patterns onto the object surface wherein the 0° reference phases for the coarse and fine fringe patterns are precisely calibrated to have a known relationship. For each projection of a coarse or fine fringe pattern, three or more images are acquired to implement the phase shift technique described above.
• a coarse fringe pattern is projected, imaged, phase shifted once, imaged again, phase shifted a second time, and imaged again to acquire a total of three images.
• the same three image procedure is implemented for the fine fringe pattern.
• a higher resolution determination of the fractional phase ⁇ f rac can be determined.
• a higher resolution determination of the phase ⁇ (a "fine ⁇ ") can be determined using the coarse ⁇ and the fractional phase ⁇ f mc .
• the coarse phase ⁇ is used to "unwrap" the fringe number for the fractional phase ⁇ f mc to permit determination of the fine phase ⁇ .
• a higher resolution determination of the distance z P is calculated based on the fine phase ⁇ .
• the foregoing unwrapping procedure is performed for each pixel to determine respective x, y and z coordinates and to generate a comprehensive higher resolution profile, or map, of the surface contour.
• the iterative unwrapping technique permits the phase ⁇ to be determined with sufficient resolution, generally it is difficult to generate coarse fringe patterns in which only one central fringe is projected in the field of view.
• a large grating pitch d is required.
• the grating pitch d is increased, the grating generally becomes less efficient, that is, less optical power is diffracted into the non-zero order modes. Consequently, the available laser power can limit the grating pitch d if sufficient laser power is not otherwise present in the first order modes to generate fringe pattern intensities sufficient for imaging.
• One conventional technique for simulating a coarse fringe pattern employs two less coarse (i.e., higher spatial frequency) diffraction gratings having slightly different grating pitches d. Consequently, the fringe patterns generated by such gratings have similar but not identical spatial frequencies. Images of the two fringe patterns are analyzed to determine a beat frequency between the similar spatial frequencies to synthesize a significantly coarser fringe pattern for the determination of the phase ⁇ without any 2 ⁇ ambiguity. Although the synthesized coarse fringe pattern requires an additional diffraction grating, the optical power efficiency limitation typically associated with a single coarse diffraction grating is eliminated.
• the diffraction gratings are sequentially moved into the path of a laser beam to generate the different fringe patterns.
• each diffraction grating generates a set of diffracted beams which are focused by an objective and spatially filtered to isolate the pair of first order beams to generate the fringe pattern.
• the distance between the apertures of the spatial filter and the size of the apertures are selected to accommodate the different source separation distances (e.g., a.f me and a coarse ) between the focused first order beams.
• a single channel interferometer projector is implemented in which different diffraction gratings are multiplexed into the beam path to generate different fringe patterns.
• One aspect of the present invention is directed to a multiple channel interferometer projector that can be employed as a replacement for the projector 14 in the measurement system 10 depicted in FIG. 1 and FIG. 2.
• the multiple channel interferometer projector includes two or more interferometer channels that are spatially separated from each other.
• the multiple channel interferometer projector achieves a significant improvement in system performance and measurement accuracy in comparison to the single channel interferometer projectors discussed above.
• the need to alternately move different diffraction gratings into the coherent beam path is eliminated.
• the only grating movement is a significantly smaller grating displacement for each diffraction grating relative to a reference phase position to implement the phase shifting technique described above.
• the major reduction in the grating displacement yields a significant decrease in the mechanical complexity, weight and size of the projector, an improvement in overall projector stability, and a significant increase in measurement speed and accuracy.
• Another advantage is the higher system reliability provided by the projector. If one of the interferometer channels fails to operate properly, surface contour measurements may still be possible using the other channels. Moreover, in some instances, if data obtained using one of the channels is inconsistent with data obtained from other channels, the suspect data can be ignored and the user can be alerted to the discrepancy.
• the multiple, spatially diverse interferometer channels provide additional advantages. Averaging the measurements obtained from multiple channels results in a decrease in the effect of optical noise on the calculated surface contour data. For example, measurement error introduced by photon shot noise and speckle decreases by an amount proportional to the square root of the number of channels averaged. In particular, speckle error is significantly reduced by analyzing multiple images having uncorrelated speckle due to the spatial separation and associated angle diversity between the two fine interferometer channels. As a result, the z coordinates of an imaged surface contour can be determined with high accuracy over a large field of view. In various embodiments described below, the z coordinates can be determined with an accuracy of better than 10 micrometers over a 0,5 meter field of view.
• a further advantage of multiple channels is the reduction of the effects of projector optical aberrations on measurement accuracy.
• Each interferometer channel exhibits inaccuracies due to specific optical aberrations caused by the optical components in the channel.
• multiple channel measurements are averaged together to generate a final measurement result and the effect of individual channel aberration errors are decreased due to the uncorrelated nature of the aberrations between channels.
• the spatial frequency of the fine fringe patterns can be reduced in many applications while maintaining sufficient measurement resolution.
• utilizing fine fringe patterns of high spatial frequency for increased measurement resolution can limit the practical imaging depth of field of the camera, therefore there can be a compromise required between measurement resolution and imaging depth of field.
• the substantial noise reduction in the measurement process achieved with the multiple channel interferometer of the present invention can, in many instances, provide a significant improvement in signal-to-noise ratio to allow a reduction in the spatial frequency of the fine fringe patterns without sacrificing a desired measurement resolution or a greater imaging distance between the camera and the object surface.
• FIG. 5 A and FIG. 5B illustrate front and rear perspective views, respectively, of a multiple channel interferometer projector 70 constructed in accordance with the present invention.
• the projector 70 has an approximately cylindrical form, a length of approximately 140 millimeters, a diameter of approximately 65 millimeters and a weight of less than 1 kilogram.
• the projector 70 includes four interferometer channels although in other embodiments different numbers of channels are employed.
• the projector 70 includes a front end cap 74 having four passageways 78A, 78B, 78C and 78D (generally 78) through which pass the optical radiation from the interferometer channels.
• a rear end cap 82 includes four optical couplers 86A, 86B, 86C and 86D (generally 86) to provide for coupling an optical light source to each of the four interferometer channels.
• the interferometer channels can be sourced by a single laser beam that is multiplexed in space and optionally in time.
• a single laser source is separated into four optical signals and each optical signal is coupled to a respective one of the four optical couplers 86, for example, by an optical fiber.
• one or more interferometer channels are sourced by a dedicated laser source such that at least two laser sources are utilized with the projector 70.
• a laser source used for the multiple channel interferometer projector 70 of the present disclosure has a wavelength of 658 nanometers and generates approximately 50 to 100 milliwatts of optical power.
• different interferometer channels of the projector 70 utilize different wavelengths of radiation.
• the fringe patterns generated by two or more spatially-separated different- wavelength interferometer channels are projected simultaneously onto an object surface, imaged simultaneously by a camera, and processed separately according to the wavelengths.
• image data generated at the camera can be separated according to color and processed independently.
• fringe patterns projected by two or more spatially-separated different-wavelength can be temporally multiplexed and synchronized with data acquisition by the camera for separate processing.
• FIG. 6A and FIG. 6B illustrate front and rear exploded perspective views, respectively, of the projector 70 shown in FIG. 5 A and FIG. 5B.
• each module includes one or more components for each interferometer channel so that a total of four component sets are employed per module.
• Each component set of a given module comprises a particular functional portion of a respective one of the four interferometer channels.
• the projector 70 includes a collimator lens module 90 including four fiber couplers 94A, 94B, 94C and 94D and four collimator lenses 98A 5 98B, 98C and 98D (generally 98).
• Each coupler-lens pair receives laser light in a respective interferometer channel.
• the laser light is supplied to the projector 70 through optical fibers secured by the couplers 86 of the rear end cap 82.
• the collimator lens module 90 is configured such that an end face of the optical fiber for each channel is aligned to a corresponding collimator lens 98.
• the projector 70 also includes a diffraction grating module 102 that includes four diffraction gratings, hi one embodiment, the diffraction gratings are integral to a common substrate disposed over four clear apertures 108 A, 108B, 108C and 108D so that each grating is irradiated with collimated laser light from a corresponding lens 98 in the collimator lens module 90.
• the four diffraction gratings, HOB, HOC and 110 D are individually fabricated and then attached to a common platform 106 as shown in FIG. 7.
• the module 102 has an approximate diameter D of 67 millimeters and an approximate thickness T of 19 millimeters.
• the common substrate 106 (or platform) can be laterally displaced along a direction perpendicular to the central axis of the projector 70 such that the fringe patterns projected on an object surface are translated along a direction parallel to the x-axis of the coordinate system 30 (see FIG. 3).
• the diffraction grating module 102 can include a single axis translation stage coupled to the common substrate 106 (or platform), one or more actuators (e.g., piezoelectric transducers) and one or more high accuracy position sensors (e.g., capacitive probes) to provide closed-loop precision shifting (i.e., "micro-positioning") for implementing the phase shift technique described above.
• actuators e.g., piezoelectric transducers
• high accuracy position sensors e.g., capacitive probes
• the first diffraction orders generated by the diffraction gratings 110 are used to generate the two sources of radiation used for each interferometer channel.
• the diffraction gratings 110 are transmissive optical components configured for high efficiency of the first diffraction orders.
• the diffraction gratings are configured as 50% duty cycle binary phase gratings having a phase depth of one-half the wavelength ⁇ of the laser light or, equivalently, the physical depth ⁇ of the grating thickness variation is ⁇ /2(n-l) where n is the index of refraction of the grating substrate 106.
• Diffraction gratings fabricated with these properties can transmit approximately 81% of the incident laser light in the first order beams.
• the diffraction gratings 110 are fabricated on a fused silica substrate having an index of refraction n of 1.457 at a wavelength ⁇ of 658 nanometers.
• the physical depth ⁇ of the gratings is 0.72 micrometers and the first order beams include approximately 79% of the optical power in the incident laser beam.
• the diffraction gratings 110 are fabricated on a fused silica substrate having a thickness of approximately 2.5 millimeters. Four grating patterns are written on a photomask covering the substrate using electron beam lithography and ion etching is used to form the patterns in the fused silica. Each diffraction grating 110 is rectangular with dimensions of approximately 8 millimeters by 10 millimeters.
• both sides of the grating substrate are anti-reflection (AR) coated.
• AR anti-reflection
• a multi-layer dielectric AR coating can be applied to limit the reflectivity per surface to be less than 0.5% at the operating wavelength ⁇ .
• FIG. 8 A and FIG. 8B illustrate detailed rear and front views, respectively, of the diffraction grating module 102.
• the grating substrate is not shown so that other components of the module 102 can be viewed.
• a centrally located translation stage 150 provides movement along a single axis parallel to the bottom flat edge of the module 102.
• the translation stage 150 includes small threaded bores 174 for attachment of the grating substrate 106 using screws. Alternatively, any of a variety of adhesives can be used to attach the grating substrate 106.
• the movement of the stage 150 is facilitated by two symmetrically disposed flexure assemblies 154.
• An electrically controlled actuator 158 e.g., a piezoelectric transducer coupled to the translation stage 150 provides precise linear motion by exerting a force on the stage 150 that deflects the flexure assemblies 154.
• a precision position sensor 162 such as a capacitive probe is coupled to the stage 150 to detect motion.
• the actuator 158 and the position sensor 162 are part of a closed-loop servo control system for positioning the stage 150.
• the actuator 158 is controlled by electrical signals transmitted over wires 164 and a feedback signal from the position sensor 162 is transmitted over cable 166.
• a wire 170 is coupled to a reference voltage (i.e., ground, not shown) and a ground plane of the module 102.
• the translation stage 150 and control components are configured to facilitate stage movement over a range of approximately 250 micrometers, with a positional resolution of approximately 30 nanometers.
• a translation stage assembly and associated control components similar to those described above are available from Dynamic Structures and Materials, LLC, of Franklin, Tennessee.
• At least one of the actuator 158 and the position sensor 162 are coupled to a processor (e.g., processor 30 in FIG. 1) through the wires 164 and cable 166), and the processor can be configured to implement a servo control loop for precision movement of the translation stage 150.
• a processor e.g., processor 30 in FIG. 1
• Other control components can be employed in addition to. or in place of the processor to facilitate control loop signal conditioning and amplification.
• the projector 70 also includes an objective and spatial filter module 114 to optically process the laser light from the diffraction gratings.
• the module 114 includes four proj ection lenses 118 A, 118B, 118C and 118D (generally 118).
• Each lens 118 focuses the first order diffracted beams from a respective diffraction grating.
• the focal length of each lens 118 if manufactured to a tolerance to ensure that the focused beams are aligned with the apertures of a respective spatial filter.
• the focal length can vary according to implementation. In one exemplary implementation, each lens has an effective focal length of 8 millimeters.
• each projection lens determines how quickly the fringe pattern expands as it exist the projector 70 and, therefore, can be selected to accommodate the required standoff distance between the projector 70 and the object surface.
• the numerical aperture is 0.45 and the angle of the first order beams with respect to the channel axis is approximately ⁇ 0.2°.
• Each projection lens 118 is adapted to operate over a finite wavelength range about the wavelength ⁇ for a given channel (e.g., 20 nanometers) to accommodate small changes in wavelength, for example, due to changes in operating temperature.
• each projection lens 118 has an AR coating such as a multiple-layer dielectric coating for improved optical transmission.
• Each diffraction grating is positioned relative to a respective lens 118 such that the plane of the diffraction grating is imaged onto the object surface. Consequently, any error in the pointing angle of the laser illumination of a grating does not result in a significant change in the position of the No fringe location at the object plane.
• the objective and spatial filter module 114 also includes four spatial filters 122 A, 122B, 122C and 122D (generally 122).
• Each spatial filter 122 includes two spaced apart pinhole apertures to pass the focused beams of the first diffracted orders.
• each spatial filter 122 is constructed of a thin fused silica substrate coated with an opaque metal layer such as a chromium layer. The metal layer is etched using a photolithographic technique to create two transparent pinholes. The fused silica substrate has an AR coating on both sides for improved transmission through the pinhole apertures.
• the spatial filters 122 provide a significant advantage over spatial filters employed in conventional interferometer projectors. More specifically, conventional projectors utilize a spatial filter having multiple pairs of pinhole apertures with each pair matched to a particular diffraction grating. Thus only one pair of pinhole apertures provides the desired beams. Other pairs of pinhole apertures can pass coherent light which can create other less intense fringe patterns having different spatial frequencies on the object surface. As a consequence, the surface contour data is less accurate and can exhibit harmonic variations.
• optical processing methods and devices compensate for intensity variations in the fringe patterns due to varying surface reflectivity.
• the compensation is based on modifying the intensity profile of a fringe pattern (i.e., "intensity shaping") at the projector.
• the projector 70' includes an. intensity shaping module 124 disposed between the front end cap 74 and the objective and spatial filter module 114 as shown in FIG. 9.
• the intensity shaping module 124 includes four optical processors 128A, 128B, 128C and 128D (generally 128). In an alternative embodiment, the intensity shaping module 124 includes a single optical processor that performs optical processing for two or more channels.
• the optical processors 128 are configured to selectively attenuate portions of one or more fringe patterns projected from the focusing objective and spatial filter module 114.
• one or more of the optical processors 128 are provided as high resolution liquid crystal displays (LCDs) wherein the optical transmissivity of individual pixels of each LCD are individually controlled, for example, by a processor such as that depicted in FIG. 1. In this manner, intensity shaping of the projected fringe pattern can be utilized to compensate for a variety of complex surface reflectivities.
• LCDs liquid crystal displays
• FIG. 10 is a cross-sectional block diagram depicting a crossbar 126 for rigidly coupling a camera 22 to a projector 14 according to an embodiment of the invention and FIG. 11 shows a perspective view of an assembled crossbar 126.
• the crossbar 126 is configured as a sleeve-like structure that includes an outer tube 130 coupled to an inner tube 134 through two O-rings 138.
• the O-rings 138 are located along the length of the inner tube 134 at two positions determined to significantly reduce changes in the orientation of the camera 22 relative to the projector 14 and to significantly reduce changes in the translation of the camera 22 and the projector 14 relative to a mounting plate 142 secured to the outer tube 130.
• the changes are due primarily to gravity and typically occur when the crossbar 126 is reoriented.
• the body weight of the inner tube 134 compensates for angular changes caused by the two loads (i.e., the camera 22 and the projector 14).
• the O-rings 138 reduce local stresses and distribute the weight of the inner tube 134 over a significant surface area of the inner surface of the outer tube 130.
• the inner tube 134 is kinematically constrained for axial and radial movement.
• the outer tube 130, inner tube 134 and mounting plate 142 are fabricated from identical materials or materials for which the coefficients of thermal expansion are closely matched so that thermal changes in the measurement environment do not significantly affect measurement accuracy.
• the outer tube 130, inner tube 134 and mounting plate 142 are fabricated from aluminum.
• each pair of coherent light sources is spatially separated from the other pairs of coherent light sources in the projector.
• methods for fringe unwrapping and surface contour data generation for a multiple channel interferometer projector in accordance with the invention take into consideration the geometric diversity of the coherent light sources as describe in detail below.
• the spatial frequency of the fringe pattern for each channel is selected according to particular linear combinations of multiple fringe patterns to facilitate phase computations that accommodate fringe unwrapping. Using generalized relationships for the generation of different fringe patterns based on an arbitrary number of channels, comprehensive characterization of surface contour measurements using multiple interferometer channels is possible.
• the spatial frequency of a projected fringe pattern is determined by the source separation distance a for the respective channel.
• the source separation distance a is related to the grating pitch d of the diffraction grating employed in the channel according to the relationship
• phase ⁇ for a given point of interest P on an obj ect surface is related to the rational quantity N given in Eq. (1) which represents the total integer and fractional number of fringes between a point of interest and a reference point.
• FIG. 12 shows the image 62 of FIG. 3 where phase is expressed in terms of the fringe number N. From Eq. (1), for a fringe pattern based on a source separation distance ⁇ ,-, the fringe number Ni associated with a point of interest P is given by
• N QNI J I and N ( F RAQI are the integer and fractional components of Ni, respectively.
• the value o ⁇ N ( i N ⁇ is zero for the peak 104 of the central fringe N 0 and the separation between neighboring peaks in the fringe pattern is one.
• N (F1U Qt represents the distance from the peak intensity of a given fringe; accordingly, N ( F R A QI can have a value between -0.5 and 0.5.
• N ( F R A QI can have a value between -0.5 and 0.5.
• the fractional portion N (FMQ i of the fringe number N t for the point of interest P can be determined with any certainty. Accordingly, the fractional portion Npju Q t serves as the basis for constructing phase maps and ultimately a profile of the imaged surface contour.
• a value ⁇ is defined by
• N (FRAC ) I for the C channels are determined from respective images, weighted by an integer fo, and summed to obtain the value ⁇ .
• Eq. (6) represents a synthesized fringe pattern that is a linear combination of the projected fringe patterns where the value ⁇ represents a fractional portion N (FMC)synth of a synthesized fringe number N syn th- Consequently, a synthesized source separation distance a syn th is given by
• /c,- and ⁇ ,- are selected based on the number C of available channels such that -0.5 ⁇ ⁇ ⁇ +0.5 to ensure that there is no integer portion of the quantity N synth associated with the synthesized fringe pattern, hi this manner, the quantity (x/R) is determined directly from ⁇ (i.e., without 2 ⁇ ambiguity) as defined by Eq. (4) based on measurements of N (FMC)I from acquired images according to the relation
• condition ⁇ ⁇ 0.5 represents a synthesized coarse fringe pattern based on a linear combination of finer fringe patterns.
• (x/R) can be determined unambiguously from Eq. (8), the resolution is lower than if the determination of (x/R) is based on Eq. (5). Accordingly, the result of (x/R) as determined by Eq. (8) can be used to determine the integer portion Af ⁇ r ⁇ ,- according to Eq. (5) to complete the fringe unwrapping analysis. More specifically, ⁇ f ⁇ vzj; can be expressed as
• N(I N T)I is determined for each channel
• Eq. (5) is evaluated again for each channel using the measured N ( PRA C)I to determine a more accurate result for the quantity (x/R).
• the higher resolution values for (x/R) can be averaged together to mitigate the effects of optical noise and significantly improve the signal-to-noise ratio.
• a dynamic range ⁇ for the quantity (x/R) can be defined such that - ⁇ ⁇ — ⁇ + ⁇
• values of Iq and a t can be selected from Eq. (7) to satisfy Eq. (10).
• an operating wavelength ⁇ of 658 nanometers and a dynamic range ⁇ of 0.25 results in a synthesized source separation distance a synth that is less than 1.316 micrometers.
• the synthesized source separation distance a synt h is 0.6 micrometers.
• the three different fringe patterns have similar source separation distances aj, a ⁇ and a 3 and generate fine fringe patterns such that the quantity (x/R) derived from Eq. (5) using unwrapped values for N(iN ⁇ has sufficiently high resolution.
• the values for N(FMQi, N( FM C)2 and N(FRAC)3 measured in the acquired images are based on fringe patterns with high intensity slopes to provide for higher resolution in the determination of z coordinates. Additionally, photon shot noise and speckle noise are significantly reduced due to the averaging of the multiple measurements.
• Measurement noise is also addressed to ensure that no unwrapping errors are made in the determination o ⁇ N ( m ⁇ )i according to Eq. (9). More specifically, every measurement OfN (FRAQt from an acquired image includes a noise component ⁇ ,- expressed in terms of a fractional fringe number. When determining ⁇ according to Eq. (4), any noise component ⁇ , in a given measurement is multiplied by a corresponding fa. Therefore, an initial determination of (x/R) according to the synthesized coarse fringe pattern technique generally includes multiple noise components. Eq. (4) can be modified as follows to include these noise components
• Eq. (16) represents a Diophantine equation (i.e., an equation in which only integer solutions are allowed) which can be solved using known mathematical techniques.
• Eq. (4) yields a unique integer value for ⁇ provided that kj and ⁇ , are appropriately selected such that the condition described by Eq. (15) is satisfied.
• the integer value for ⁇ then can be used according to Eq. (16) to unambiguously determine the values of N (INT) I, which then can be added to the corresponding measured values of ' N F R AC ® to determine the fringe numbers N) according to Eq. (3).
• a value for the quantity (x/R) can be determined for each N,- (i.e., for each channel) with high resolution, and these values are utilized in a triangulation process in which information from the multiple channels is averaged to determine the profile of a surface contour with high precision.
• Eq. (11) can be expressed as (22) where the term (m; + m ⁇ ) ⁇ represents the noise component and, as described above, a similar per channel noise budget ⁇ is assumed for each of the two fringe patterns. In effect, mi and M2 act as noise multipliers.
• nij, m. 2 , ni 3 and m 4 are monotonically increasing relatively prime integers, such that:
• the system of three fringe patterns is treated as two subsystems each having two fringe patterns and sharing a common fringe pattern, namely, a first subsystem based on aj and a 2 , and a second subsystem based on m and «?.
• Each subsystem is treated separately to satisfy the relationships given in Eq. (15) and Eq. (16).
• Eq. (18) can therefore be expressed as follows in terms of (Tj 2 (to indicate the subsystem of fringe patterns based on aj and «2)-'
• the first fringe subsystem based on aj and « 2 can be used to "unwrap" N(J MT > 2 , modulo ni2-
• the second fringe subsystem based on # 2 and a ⁇ can be used to "unwrap" N (J N T ) 2 , modulo m?.
• the value N(m ⁇ )2 can be uniquely determined up to a value corresponding to the product of m. 2 and /r ⁇ j.
• a per channel noise budget ⁇ can be established to ensure that no unwrapping errors can occur for the determination of N (JNT)2 - ⁇ ? I , 2 and 02,3 should be measured and rounded correctly for proper evaluation of Eq. (25) and Eq. (28), respectively.
• Eq. (34) mi, ni2, ni3 and iri4 are monotonically increasing integers, therefore the per channel noise budget expressed in Eq. (34) is more conservative than that given by Eq. (23). Accordingly, Eq. (34) is used to establish the noise budget for the three fringe pattern system. However, if the w-values for the three fringe pattern system are significantly smaller than those for the two fringe pattern system to achieve the same dynamic range as discussed above, the noise budget for the three fringe pattern system can be appreciably higher than that of the two fringe pattern system. Consequently, the three fringe pattern system offers appreciable benefits over the two fringe pattern system by affording an increased noise budget for the same dynamic range.
• N (INT)2 is uniquely determined, a value for (x/R) is determined for the fringe pattern based on a 2 (Eq. (3)) and used to unwrap N ( I NT ) I and N (INT )3 based on Eq. (9).
• N (IN i )I is determined for each channel, Eq. (5) can be evaluated again based on the originally measured N (FRAC) ⁇ to generate more accurate results for the quantity (x/R) for each channel.
• FRAC originally measured N
• fine fringe patterns of approximately 70 to 80 fringes are selected for inclusion in the images to achieve adequate resolution.
• the per channel noise budget is 0.026 fringes.
• the determination of a grating pitch d 2 for the second channel represents a second design selection.
• the grating pitch d 2 should ensure a high transmission efficiency for the first diffraction orders; accordingly a grating pitch d 2 of 200 micrometers is chosen.
• a wavelength A of 658 nanometers and an effective focal length/of 8.00 millimeters the source separation distance a 2 is 52.64 micrometers.
• the common base distance a is 6.58 micrometers and aj 46.06 micrometers.
• the grating pitch dj can be determined from Eq. (2) to be 228.571 micrometers.
• the common base distance/? is 5.849 micrometers
• the source separation distance 0, 3 is 58.489 micrometers
• the grating pitch d$ is 180 micrometers.
• Table 2 summarizes the grating pitches and corresponding source separation distances for the above example of a three channel system.
• the three fringe patterns have similar source separation distances ai, a.2 and «3 so as to generate fine fringe patterns for which the quantity (x/R) derived from Eq. (5) using unwrapped values for N(wi)i has sufficient resolution.
• the respective N(FRAQI, N(F R AQ2 a *id N(FRAC)3 measured in the acquired images are based on fringe patterns having large variations in intensity over small image plane distances to achieve higher resolution in the determination of z coordinates. Additionally, photon shot noise and speckle noise are significantly reduced due to the averaging of the multiple measurements.
• the fringe number Ni is referred to as an ideal fringe number N 1 -, , - ⁇ 0 / if the assumption is true, hi the various embodiments of the multiple channel interferometer projector described above, the assumption is not uniformly valid for all channels as each channel is spatially offset from the other channels.
• the real fringe number Ni, rea ⁇ for each channel is given by
• the offsets are in the x and y directions and any unintended offset in the z direction is typically small (e.g., less than 100 micrometers).
• FIG. 13 schematically depicts an end view of a three channel interferometer projector 14'.
• the spatial offsets of the three active channels (channel 0 (CH 0), channel 1 (CH 1) and channel 2 (CH 2)) are shown.
• a fourth channel (CH 4) is inactive and is reserved for a system monitoring function or other purpose, hi the following description, channel 2 is selected as a master channel, i.e., a channel defined to be on a "collinear axis" and the channel to which the spatial offsets of the two other active channels are referenced.
• a master channel i.e., a channel defined to be on a "collinear axis" and the channel to which the spatial offsets of the two other active channels are referenced.
• ⁇ x o, ⁇ y o, ⁇ x i and ⁇ y ⁇ are the x and y offsets of channel 0 and channel 1. If the offsets are small with respect to the distance z, the real (i.e., measured) fringe numbers No ,rea i, ⁇ l . rea i and N 2 ,reai are given by
• the camera acquiring images of the fringes includes an array of pixels.
• Each pixel has a location (i, j) defined in the camera image plane and is associated with a point on the surface of the object being measured, hi a camera-based coordinate system where three dimensional coordinates of the object surface are given as (xcamera, y camera, Zcamera) and the origin is at the camera pupil location, the following coordinate relationships are valid
• PixelPitch is the spatial separation between pixels and L is the separation between the camera pupil and the image plane (i.e., the "pixel plane").
• L is the separation between the camera pupil and the image plane (i.e., the "pixel plane").
• the three dimensional locus of points mapping to the pixels can be parameterized as a function of z camem as
• the object points can be specified in terms of coordinates (x, y, z) according to the coordinate system 38 with an origin at a multiple channel interferometer projector located at a projector position as illustrated in FIG. 1 and FIG. 2 according to
• the values of the collinear fringe numbers Noj d eai and N1, ideal can be used according to the unwrapping procedure described above for collinear interferometer channels.
• the desired fringe number results N0final and N1final are determined by modifying the unwrapped collinear fringe numbers by the respective collinear adjustment value applied prior to the unwrapping as follows and
• channel 2 (CH 2), by initial selection, was defined to be on the "collinear axis.”
• FIG. 14 is a flowchart summarizing method steps described above to calculate the fringe numbers for a pixel for fringe patterns generated by a multiple channel interferometer projector according to an embodiment of the invention.
• the fringe numbers N ⁇ >rea ⁇ are determined (step 210) by measurement for each active channel.
• the measured fringe numbers are adjusted (step 220) using the collinear adjustment values to obtain collinear fringe numbers Nijde a i, i-e., fringe numbers that correspond to a system in which all the interferometer channels share the same axis.
• the collinear fringe numbers Ni, i de al are unwrapped (step 230) and then the unwrapped values are modified (step 240) to remove the prior adjustment used for collinearization.
• the method 200 advantageously permits the unwrapping of fringe numbers to occur in "collinear space” using a technique that is also valid for conventional single channel interferometers.
• the unwrapped values are then modified to "remove” the adjustments applied to the offset channels for an accurate determination of the fringe number for each channel.

Landscapes

• Physics & Mathematics (AREA)
• General Physics & Mathematics (AREA)
• Engineering & Computer Science (AREA)
• Computer Vision & Pattern Recognition (AREA)
• Length Measuring Devices By Optical Means (AREA)
• Instruments For Measurement Of Length By Optical Means (AREA)
• A Measuring Device Byusing Mechanical Method (AREA)
• Length Measuring Devices With Unspecified Measuring Means (AREA)

Applications Claiming Priority (2)

Publications (2)

Family

ID=36739915

Family Applications (4)

Family Applications Before (3)

Country Status (3)

Families Citing this family (71)

Family Cites Families (27)

• 2006
  • 2006-04-04 WO PCT/US2006/012438 patent/WO2006107928A2/en active Application Filing
  • 2006-04-04 EP EP06749214.0A patent/EP1875161B1/de active Active
  • 2006-04-04 EP EP13175579.5A patent/EP2657761A3/de not_active Withdrawn
  • 2006-04-04 EP EP06749215.7A patent/EP1875162B1/de not_active Ceased
  • 2006-04-04 US US11/910,638 patent/US7751063B2/en active Active
  • 2006-04-04 WO PCT/US2006/012480 patent/WO2006107955A1/en active Application Filing
  • 2006-04-04 WO PCT/US2006/012440 patent/WO2006107929A1/en active Application Filing
  • 2006-04-04 EP EP06749233.0A patent/EP1875163B1/de active Active
  • 2006-04-04 US US11/910,642 patent/US7595892B2/en not_active Expired - Fee Related
  • 2006-04-04 US US11/910,646 patent/US7599071B2/en not_active Expired - Fee Related
• 2009
  • 2009-09-03 US US12/553,526 patent/US20090324212A1/en not_active Abandoned

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events